96 ROSS G. HARRISON 
bud is implanted in normal location, triplicate appendages could 
be accounted for in the following way: one member derived 
directly from the grafted bud ; one member, of mirror symmetry, 
by an inverse regeneration from the base of the bud, and the 
third member, of original asymmetry, from the wound surface 
of the host. This view is, however, not borne out by the present 
experiments, i*^*^ 
When the whole of the evidence bearing on the question is 
taken into consideration, one cannot but think that too much 
weight has been placed by Bateson and his followers on the double 
supernumerary. The other class of cases, where the single 
supernumerary is symmetrical with the normal appendage 
with which it is associated, while neither so numerous nor so 
spectacular, is nevertheless of wide occurrence. Cases reported 
by Tornier ('97), Przibram ('02), Reed ('04), Zeleny ('05), and 
Megusar ('07) show that truly double appendages in mirror 
symmetry with respect to each other may be formed by constric- 
^"^ Se dunque noi considerassimo uno di questi innesti praticati invece che in 
una regione lontana (come p. es. nell'orbita), nell'immediata vicinanza della 
regione donde fu tolto I'innesto, noi osserveremmo I'uno presso dell'altro lo svil- 
uppo oltre che dell'arto normale, anche dell'arto rigenerato dalla superficie di 
sezione della regione prossimale del corpo, nonche dell'arto sviluppatosi dalla 
superficie di sezione della regione perif erica, identico all'arto che lo ha prodotto, 
ma con simmetria speculare. La identita anche di quest o fcnomeno con la dop- 
pia rigenerazione inversa dalle due superficie di una ferita risulta in questo modo 
evidente. Della Valle: op. cit. p. 125. 
There is opportunity to test this hypothesis by comparing the experiments in 
which the wound-bed was cleaned with those where it was not. In the former, 
regeneration from the host is precluded (p. 6), and triplicate limbs could only 
arise by a second reduplication from the base of the graft; whereas in the latter, 
regeneration from the host should occur in a large number of cases, if at all, and 
thus yield a large proportion of triplicate appendages. An examination of the 
results shows that this is not the case. In the first place, as shown in table 2, 
the total number of reduplications in the series with cleaned wounds is fifty-three, 
which is 56 per cent of the total number of positive experiments, while there are 
but sixteen cases (33.3 per cent) in the group with non-cleaned wounds. The 
disproportion is much greater when the number of triplicate appendages in each 
group is compared. Out of a total of eighty-seven cases old enough to be deter- 
mined, there are twenty-five triplicate limbs (28.7 per cent) in the clean-wound 
experiments and only three in forty-eight cases (6.25 per cent) in the others. 
It is quite clear, then, that leaving in the wound-bed cells that are capable of 
giving rise to a new limb reduces greatly, instead of increasing, the chance of 
formation of supernumerary limbs, so that Della Valle's suggestion is untenable. 
